Despite surpassing human performance in certain tasks, deep learning models face several challenges, such as large model parameters, high computational resource consumption, and high data quality requirements. The human brain is currently the only true intelligent system, with extremely low energy consumption and various cognitive functions. Clearly, learning from the brain's information-processing mechanisms to develop more powerful and versatile machine intelligence is very promising. Spiking neural networks, based on the brain's "spiking" computational framework, can reduce computing platforms' energy consumption while achieving artificial intelligence. This forum invites four academic experts to delve into the structure, models, and learning theory methods of spiking neural networks, as well as brain-inspired theories and applications of visual cognition, to promote academic exchange in the field of brain-inspired intelligence.
尽管深度学习模型已经在在某些任务中胜过人类，但面临模型参数大，计算资源消耗高和数据质量要求高等问题。人类大脑是目前唯一真正的智能系统，能耗极低，且具有的不同认知功能。显然，学习大脑的信息处理机制，去建立更强大和更通用的机器智能是非常有前景的。基于人脑的“脉冲”模拟计算框架下的脉冲神经网络有望在实现人工智能的同时，降低计算平台的能耗。本论坛邀请四位学术界专家，深入探讨脉冲神经网络结构、模型和学习理论方法，以及大脑仿生视觉理论和应用，促进脑启发智能领域学术交流。